friendlyarm / install-opencv-on-friendlycore Goto Github PK

Script for installing OpenCV on FriendlyCore.

Shell 100.00%

install-opencv-on-friendlycore's Introduction

Installing OpenCV 4.2.0 on FriendlyCore/FriendlyDesktop

The easiest way to install it is to run FriendlyELEC's script.
Here are the packages and utilities your system will have after you follow the instructions in this tutorial:

Qt 5.10.0 version of the HighGUI module (Better 2D window interface with zoom, image saving capabilities, etc)
C++ interface and examples
C interface and examples
Python 3.6+ interface and examples

Currently supported boards

RK3399
NanoPC T4
NanoPC M4
NanoPC NEO4
Som-RK3399

If you are using the s5p4418/s5p6818 development board, please checkout to the "s5pxx18" branch.

Installation

Note: OpenCV has been pre-installed in FriendlyCore/FriendlyDesktop (Version after 201905) and does not require manual installation.
Please download the latest FriendlyCore/FriendlyDesktop Image file from the following URL: http://download.friendlyarm.com

To make it easy to test python3 examples, you may copy cv-env.sh to the system directory，run the commands below:

git clone https://github.com/friendlyarm/install-opencv-on-friendlycore
cd install-opencv-on-friendlycore
cp examples/py/cv-env.sh /usr/bin/

Test out the OpenCV 4.2.0 and Python3 install

Run the commands below:

cd ~/install-opencv-on-friendlycore/examples/py/
. cv-env.sh
python ver.py

Will display:

4.2.0

it will activate a virtualenv, if you want to switch projects or otherwise leave your virtualenv, simply run:

deactivate

Test python sample included in OpenCV 4.2.0: turing

. ~/install-opencv-on-friendlycore/cv-env.sh
. setqt5env
cd /usr/local/share/opencv4/samples/python
python turing.py

Build Qt5 MIPI-Camera example

Note: To run this demo you will need a camera(ov13850/ov4689/logitech c920 pro) and a display connected.
Run the commands below:

cd ~/install-opencv-on-friendlycore/examples/qt5/mipi-camera-videoprocessor/
qmake-qt5 .
make -j4

Run on FriendlyDesktop:

su pi
. setqt5env
export DISPLAY=:0.0
./mipi-camera-videoprocessor

Run on FriendlyCore:

./run-on-friendlycore.sh

Python: Stereo match example

Run the commands below:

cd ~/install-opencv-on-friendlycore/examples/py/single-camera-python-demo/
./run.sh

Python: Test single camera example

Note: To run this demo you will need a camera(ov13850/ov4689/logitech c920 pro) and a display connected.
Run the commands below:

cd ~/install-opencv-on-friendlycore/examples/py/single-camera-python-demo/
./run.sh

Python: Test dual camera example

Note: To run this demo you will need two cameras(ov13850/ov4689/logitech c920 pro) and a display connected.
Run the commands below:

cd ~/install-opencv-on-friendlycore/examples/py/dual-camera-python-demo/
./run.sh

C++: Test UVC camera example

Note: To run this demo you will need a usb camera(logitech c920 pro) and a display connected.
Build and run the commands below:

git clone https://github.com/friendlyarm/rk3399-opencv-uvccamera-cpp-demo.git
cd rk3399-opencv-uvccamera-cpp-demo
mkdir build
cd build
cmake ..
make
su pi
export DISPLAY=:0.0
. setqt5env
./opencamera

install-opencv-on-friendlycore's People

Contributors

Stargazers

Watchers

install-opencv-on-friendlycore's Issues

' Building wheel for numpy (setup.py) ... \ '

Run install-opencv.sh always stay in 'Building wheel for numpy (setup.py) ... '

OpenCv cap(0) . not working ?

in c++ VideoCapture cap(0);

gives error:

VIDIOC_QUERYCTRL: Inappropriate ioctl for device

any help ?

dependencies broken

hi,
Dependencies seems to be broken after apt-get upgrade :

#./mipi-camera-videoprocessor 
#./mipi-camera-videoprocessor: symbol lookup error: /usr/local/Trolltech/Qt-5.10.0-rk64one/plugins/egldeviceintegrations/libqeglfs-kms-integration.so: undefined symbol: gbm_create_device

neo 2 plus can used this script ？

How to capture video with OpenCV from CAM1320 module

I would like to access the CAM 1320 from Python via the openCV library. Can somebody help ? Documentation is a bit sparse in the direction...

Installing opencv on smart4418 for every usb camera.

Hi
Thanks for your valuable repository.
Is there any way to install python opencv on board smart4418 that works for every usb camera?
Regard

error when videocapture from gstreamer

when I running a demo script to capture rtsp stream using cv2.VideoCapture() in python on the fire3 as following:

if name == "main":
print("OpenCV version: {}".format(cv2.version))
gst_str = "rtspsrc location=rtsp://192.168.0.101:554/Streaming/Channels/101 protocols=GST_RTSP_LOWER_TRANS_TCP ! rtph264depay ! h264parse ! nxvideodec ! video/x-raw, width=1280, height=720, format=(string)BGRx ! appsink"
cap=cv2.VideoCapture(gst_str, cv2.CAP_GSTREAMER)

the error occurred and the output recorded as,

OpenCV version: 3.4.2-dev
mime type = video/x-h264
NumSps = 1, NumPps = 1, type = avcC type
node found for device nx-vpu-dec: /dev/video15
<<<<<<<<<< InitializeCodaVpu(Min=7, 1280x720) (ret = 0) >>>>>>>>>

(python:1011): GStreamer-CRITICAL **: gst_element_get_state: assertion 'GST_IS_ELEMENT (element)' failed

How it run without virtualenv?

Some libs do not support virtual env. How I can install CV without it?

Installing snap

I am trying to use snap on friendlycore OS on my NanoPi NEO.
I have install snapd with the commend sudo apt install snapd.

In the installation I get the following errors:

md5sum: /etc/apparmor.d/usr.lib.snapd.snap-confine: No such file or directory
snapd.failure.service is a disabled or a static unit, not starting it.
snapd.snap-repair.service is a disabled or a static unit, not starting it.
Update rpimonitor Packages Status
E: Could not get lock /var/lib/dpkg/lock-frontend - open (11: Resource temporarily unavailable)
E: Unable to acquire the dpkg frontend lock (/var/lib/dpkg/lock-frontend), is another process using it?
W: Could not lock the cache file; this usually means that dpkg or another apt tool is already installing packages. Opening in read-only mode; any changes you make to the states of packages will NOT be preserved!

And then it recognize the snap commend but when I run sudo snap install hello-world I get the following error:

error: system does not fully support snapd: cannot mount squashfs image using "squashfs": mount:
/tmp/sanity-mountpoint-573191352: mount failed: Unknown error -1

This OS is based on Ubuntu core so I was expecting it will support snap, It would be nice if you will add the support out of the box.

how to run dnn_objdetect from samples?

I'm having trouble running dnn objdetect, can please show me how to do it?

Problem with configuring the UVC resolution by cv::VideoCapture::set()

The default resolution is 640*480.
When I try to use "cv::VideoCapture::set(CAP_PROP_FRAME_WIDTH ,1280)" to change the resolution, it doesn't work.